PETALING JAYA: The latest advancements in artificial intelligence (AI) innovation are expected to be driven by small language models and agentic AI, according to IBM Malaysia.
IBM Malaysia Country general manager and technology leader Dickson Woo said SLMs offer more efficient and cost-effective AI solutions for enterprises, while agentic AI enables systems to operate more autonomously.
“It also helps in making complex decisions and carrying out tasks with little or no human intervention.
“Organisations do not need a massive, all-knowing AI model but rather an AI that is open, targeted, and efficient,” he said during the “AI ROI: Unlocking Value and Scaling Impact” presentation on Wednesday.
Woo said IBM has introduced IBM Granite, a small AI model with eight billion and two billion parameters, optimised for efficiency and performance.
“This model is integrated into the IBM watsonx platform, allowing businesses to implement AI at scale with greater transparency and collaboration within the AI community. With increased investment in open-source AI and innovations like Agentic AI, companies are expected to accelerate their digital transformation to compete in an increasingly complex global market.”
Woo noted that agentic AI, built on SLM technology, is becoming more accessible for small and medium enterprises due to its lower cost and ease of implementation.
“It is more cost-effective, and adoption is also easier from a technical standpoint,” he said, adding that the models require fewer computing resources and significantly reduce the total cost of ownership.
As these AI models consume less energy and require fewer technical resources, Woo said, SMEs can now deploy AI-driven solutions without the heavy financial burden traditionally associated with AI adoption.
Companies worldwide are increasingly turning to open-source AI tools to maximise return on investment (ROI) and drive innovation, according to IBM’s latest report, he noted.
“A study conducted by IBM in collaboration with Morning Consult and Lopez Research found that 89% of organisations plan to increase or maintain their AI investments in 2025. Of this, 62% will increase their allocations, with nearly two-fifths (39%) planning hikes of between 25% and 50%.
“Companies are now also focusing on strategic AI investments, particularly in IT operations (63%), data quality management (46%), and product and service innovation (41%),” he said.
“The study also revealed a significant rise in open-source AI adoption, with 51% of companies using it achieving positive ROI, compared to just 41% among organisations that do not take this approach,” he added.
Woo said key strategies planned by companies for 2025 include the adoption of managed cloud services (51%), recruitment of specialised AI talent (48%), and increased use of open-source solutions (48%).
Furthermore, since the introduction of IBM Granite, interest in agentic AI has grown across various industries such as healthcare, manufacturing, finance, utilities, and public services, he pointed out.
“Many organisations are already engaging with us to develop agentic AI models, and we are currently in the proof-of-concept stage with several of them,” Woo said.
